Make TableCollections iterable

[hyanwong]

No test cases yet, I'll write them once the idea is approved

[hyanwong]

So now instead of @jeromekelleher 's ts identify checking code it's now possible to do

assert all(t1==t2 for t1, t2 in zip(ts1.tables, ts2.tables) if t1[0] != 'provenances')

Which I think is a little cleaner. Or possibly, to avoid string definitions:

assert all(t1==t2 for t1, t2 in zip(ts1.tables, ts2.tables) if not isinstance(t1[1], tskit.ProvenanceTable))

[hyanwong]

Even better, we can easily check identity without the provenance timestamps having to be the same:

for t1, t2 in zip(ts1.tables, ts2.tables):
 if isinstance(t1[1], tskit.ProvenanceTable):
  assert all(t1[1].record==t2[1].record)
 else:
  assert t1==t2

[codecov]

Codecov Report

Merging #244 into master will increase coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #244      +/-   ##
==========================================
+ Coverage   86.64%   86.66%   +0.01%     
==========================================
  Files          17       17              
  Lines        9437     9446       +9     
  Branches     1712     1712              
==========================================
+ Hits         8177     8186       +9     
  Misses        750      750              
  Partials      510      510

Flag	Coverage Δ
#c_tests	86.66% <100%> (+0.01%)	⬆️
#python_tests	98.25% <100%> (ø)	⬆️

Impacted Files	Coverage Δ
python/tskit/tables.py	99.79% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4d6fdf0...9fe458d. Read the comment docs.

[jeromekelleher]

I can't think of any other meaning we could have for __iter__, so this seems like a sensible way of using it. I think we should use the simple definition based on generators for production code though (can use current definition for testing).

[hyanwong]

@jeromekelleher: do you think it is also worth including the table-check-without-timestamps code in tskit? e.g.

class TableCollection():
    ...
    def equal(self, tc_to_compare, ignore_provenance_timestamps=True):
        for t1, t2 in zip(self, tc_to_compare):
            if ignore_provenance_timestamps and isinstance(t1[1], tskit.ProvenanceTable):
                if all(t1[1].record!=t2[1].record):
                    return False
            else:
                if t1 != t2
                    return False
        return True

[jeromekelleher]

@jeromekelleher: do you think it is also worth including the table-check-without-timestamps code in tskit? e.g.

No, to be honest. We're spending more time on making this stuff that we would save.

[jeromekelleher]

Looks good, thanks and ready to merge. Can you do a quick rebase to bring it up to date please?

[hyanwong]

Done